remote auroral activity detection and modeling using low ... · 1808 e. d. schmitter: remote...

Ann. Geophys., 28, 1807–1811, 2010www.ann-geophys.net/28/1807/2010/doi:10.5194/angeo-28-1807-2010© Author(s) 2010. CC Attribution 3.0 License.

AnnalesGeophysicae

AnG

eoC

omm

unic

ate

s

Remote auroral activity detection and modeling using low frequencytransmitter signal reception at a midlatitude site

E. D. Schmitter

University of Applied Sciences Osnabrueck, 49076 Osnabrueck, Germany

Received: 10 August 2010 – Accepted: 27 September 2010 – Published: 30 September 2010

Abstract. The low frequency propagation conditions alongthe path from Iceland to Germany (52◦ N 8◦ E) using theNRK/TFK 37.5 kHz transmitter (63.9◦ N 22.5◦ W) prove asan easy to monitor and reliable proxy for north auroral ac-tivity. Signal processing using wavelet decomposition al-lows for quantitative activity level estimations. Calibrationis based upon NOAA POES auroral activity data. Using anauroral oval model for the local intensity distribution of so-lar energetic particle precipitation and a wave propagationmodel ionospheric D-layer height decreases along the pathcan be derived. This in turn gives a hint to the low latitudeextension and intensity of the auroral electrojet currents thatcan be responsible for communication and power systemsfailures.

Keywords. Ionosphere (Auroral ionosphere; Modeling andforecasting; Particle precipitation)

1 Introduction

Remote detection of auroral activity using moderately elab-orate equipment is a good complement to in situ measure-ments and satellite based recordings, especially with regardto the lower ionosphere behaviour which cannot be accessed“from top” and also with regard to nearly 24/7 availabilityof data covering the same propagation range. We followand extend previous work done byCummer et al.(1996),Cummer et al.(1997), Kikuchi and Ohtani(1986), Peter etal. (2005). Their work indicated that VLF phase and am-plitude anomalies and effective D-layer reflection height de-crease is well correlated with an increase of high energetic(E ≥ 300 keV) electron precipitation during expansion ofthe equatorward boundary of the auroral oval crossing VLF

Correspondence to:E. D. Schmitter([email protected])

propagation paths. Location and intensity of the auroral elec-trojet which is coupled to this boundary are not only of geo-physical interest but important for the assessment of possiblecommunication and electric power systems failure risks. Theorganization of the paper is as follows: in Sect. 2 we de-scribe the characteristic variations of the signal amplitude ofthe NRK/TFK 37.5 kHz transmitter (63.9◦ N 22.5◦ W, L = 5,Iceland) at the receiver site (52◦ N, 8◦ E, L = 2.2, NW Ger-many) and its wavelet based analysis used for the estimationof the auroral activity level. Section 3 deals with the auroraloval model and the propagation model needed for the calcu-lation of the effective ionospheric reflection height decreasescaused by precipitations related to auroral oval expansion.The last chapter presents results and conclusions.

2 Estimating auroral activity

For getting information about auroral activity we monitor thesignal amplitude of (very) low frequency MSK (MinimumShift Keying) transmitters within the subpolar auroral zone.The Tx station with call sign JXN (Norway, 67◦ N 14◦ E)transmits only sporadically, whereas the Tx station withcall sign NRK/TFK at Grindavik, Iceland (63.9◦ N 22.5◦ W,L = 5) usually transmits 24/7 with few dropouts. Addition-ally near to this site the Leirvogur magnetometer is situatedso that we can assess the geomagnetic activity at the trans-mitter site (URL:www.raunvis.hi.is/∼halo/lrv.html) causedby variations of the location and the intensity of the auro-ral electrojet current system - often coincident with the lowlatitude boundary of high energy auroral particle precipita-tion Cummer et al.(1997). We use 2 receivers at an averagemid latitude location of (52◦ N, 8◦ E, L = 2.2) with a dis-tance of 30 km and a great circle distance to the transmit-ter of 2210 km. Comparing the signals of the two receiverswe can discard local disturbances. The receivers use coilswith a ferrite core as sensors for the horizontal magnetic field

Published by Copernicus Publications on behalf of the European Geosciences Union.

http://creativecommons.org/licenses/by/3.0/

www.raunvis.hi.is/~halo/lrv.html

1808 E. D. Schmitter: Remote auroral activity detection

4 Ernst D. Schmitter: Remote Auroral Activity Detection

Fig. 1. Top down: NOAA POES auroral activity data, DST index,horizontal component Leirvogur magnetogram (sensitive to loca-tion and intensity of the auroral electrojet current system) and Txsignal amplitude monitored at 52N 8E (sensitive to high energy par-ticle precipitation along the propagation path).

Cummer, S.A. and Bell, T.F. and Inan, U.S. and Chenette, D.L.:VLF remote sensing of high-energy auroral particle precipita-tion, J. Geophys. Res. (USA), 7477 84, 1997

Kikuchi, T., Ohtani, A.: IV. Ionospheric Disturbances; 8. Propaga-tion Disturbances of VLF Radio Waves on the Trans-Polar Paths,Journal of the Radio Research Laboratory, Vol. 33, special issueNo. 1, pp. 203-217, 1986

Lewalle, J., Tutorial on Continuous WaveletAnalysis of Experimental Data, URL:www.ecs.syr.edu/faculty/lewalle/tutor/tutor.html, SyracuseUniversity, 1995

Peter, W. B., Chevalier, M. and Inan, U. S.: Subionospheric VLFmeasurements of the effects of geomagnetic storms on the mid-latitude D-region, 11th International Ionosphere Effects Sympo-sium, A070, May 3-5, 2005, pp. 1-8, 2005

Volland, H.: Longwave Sferics Propagation within the AtmosphericWaveguide, Chapter 3 of Handbook of Atmospheric Electrody-namics, Vol. II, Volland, H. (ed.), CRC Press, 1995

Yoshida, M, Yamauchi,T., Horie,T., Hayakawa, M.: On the gener-ation mechanism of terminator times in subionospheric VLF/LFpropagation and its possible application to seismogenic effects,Nat. Hazards Earth Syst. Sci., 8, 129134, 2008

Wait, J. R.: The Ancient and Modern History of EM Ground-WavePropagation, IEEE Antennas and Propagation Magazine, Vol. 40,No. 5, pp. 7-24, Oct. 1998

Fig. 2. Monitored signal amplitude and wavelet decomposition(bottom of wavelet decomposition: period 0, top: period 1 h).

Fig. 3. Auroral oval model for activity level 9, position at 22 UT. In-dicated are the 60 and 80 degree latitude circles, the position of thegeomagnetic pole and the propagation path from (63.9N, 22.5W) to(52N 8E) as white arrow.

Fig. 4. Model calculations for ground wave + multi hop propaga-tion. Signal amplitude calculations of the 37.5 kHz Tx at the re-ceiver site (52N, 8E, great circle distance d = 2210 km) convergessufficiently with n = 5 hops.


component of the signal. The broadband preamplified signalis fed via a soundcard into a computer, where the Tx signal isextracted via Fast Fourier Transform (FFT) and further anal-ysed. Figure1 shows an example recording for the 3 days14 July 2010 to 16 July 2010 (lowest of the 4 traces). Duringthis time some minor auroral activity took place neverthe-less showing up in a disturbed night time signal clearly to beseen between the first and the second day. Variations of lo-cation and intensity of the auroral electrojet current systemclearly show up in the Leirvogur horizontal magnetogram(3. trace). The first trace shows the total polar cap powerinput in GigaWatt as estimated by the NOAA POES satel-lites (NOAA 15-19, Polar Orbiting Environmental Satellites,URL: www.swpc.noaa.gov/pmap). Auroral activity also is aproxy for global geomagnetic activity, represented here bythe Kyoto DST index (2. trace) where the drop event coin-cides with the start of the main auroral activity.

To quantify the information content of the signal record-ings for the discrimination of quiet and disturbed parts of thetime series a wavelet decomposition is done. We decided forthe Morlet wavelet decomposition (Lewalle, 1995). Com-pared to the many wavelet transforms in use this one is es-pecially suited to detect pseudo-periodicities. The Morlet isa wave packet and in frequency space Morlet transform actslike a bunch of bandpass filters moved along the time se-ries to be analyzed. Figure2 shows the signal received on 2May 2010 and below the amplitude of pseudo periods from(nearly) 0 (bottom) to 1 h (top). Besides the terminator struc-














tures we see the signature of a disturbance after 23:00 UTwith characteristic pseudo periods between 15 and 30 min.

As a measure for the total pseudo periodic activityA(UT)of the signal within a 3 h time slot starting at time UT we usethe sum of all wavelet coefficients (absolute values) withinthat slot:

A(UT) =

∑3 h

∑pseudo periods<1 h

|Morlets| (1)

To derive an index for the auroral activity, the pseudo peri-odic activity of a quiet day directly neighbouring the activedays is subtracted. Otherwise the wavelet signatures of thesunrise/set terminators would distort the result.

We model the polar activity levelac estimated from ourlow frequency signal amplitude recordings by

ac(UT) = α (A(UT)−Aquiet(UT)) (1+2 cos(z)) (2)

z is the suns zenith angle halfways at the propagation path.By the factor(1+2cos(z)) we empirically take into accountthat day over the signal is less sensitive to additional ioniza-tion effects than during night. The scale factorα dependson the details of the Morlet transform and is calibrated bycomparing results to the NOAA POES auroral activity data.

Now further analysis has the goal to extract additionally tothe estimated auroral activity levels the effective lower iono-spheric height variations along the Tx-Rx path from the data.For this we need a model of the auroral oval particle precip-itation intensity in dependence on latitude and longitude andadditionally a wave propagation model.

3 Modeling ionospheric height variations

Based on the estimated polar activity level we can calculatethe ionizing effect along a given propagation path throughthe polar and subpolar domain using an auroral oval model.We describe it in the following subchapter. Afterwards thewave propagation model is explained.

Ann. Geophys., 28, 1807–1811, 2010 www.ann-geophys.net/28/1807/2010/

www.swpc.noaa.gov/pmap

E. D. Schmitter: Remote auroral activity detection 1809













Fig. 3. Auroral oval model for activity level 9, position at 22:00 UT.Indicated are the 60 and 80 degree latitude circles, the positionof the geomagnetic pole and the propagation path from (63.9◦ N,22.5◦ W) to (52◦ N 8◦ E) as white arrow.

3.1 The auroral oval model

For an activity levelac at universal time UT the particle pre-cipitation intensity is modelled by an elliptic Gaussian ac-cording to

I (lat,long,UT,ac) = I0w2e

−w

2σ2a (3)

With the projected coordinatesx = rcos(long), y =

r sin(long), r = re cos(lat), re = 6371 km as componentsof a vectorrT

= (x,y) we define the vectorrn = RUT(r −

rmag pole) which centers the coordinates at the magnetic poleand rotates according to the universal timeUT using theusual 2-dimensional rotation matrix with UT expressed asrotation angle. w = rn

T Q rn is the bilinear form ofrn

scaling these coordinates to an elliptic figure with semiaxesa = (1590+130ac) km andb = (1270+100ac) km with thescaling matrixQ = (1/a2 0;0 1/b2) (Chen et al., 2008).Chen et al.(2008) use the natural log of the AE-index insteadof the auroral activity indexac in the linear combinations forthe semiaxesa andb.

I0 andσa are calibrated as functions of the activity suchthat the oval shapes according to the POES auroral activ-ity maps (www.swpc.noaa.gov/pmap/) are reproduced and tosatisfy:∫

polar-capI (lat,long) d� = Itotal (4)

whereItotal is the total estimated power input into the northpolar cap area provided by the NOAA POES data.Itotal inGigaWatt is related to the activity levelac according to

Itotal= eac/2 (5)













Fig. 4. Model calculations for ground wave + multi hop propa-gation. Signal amplitude calculations of the 37.5 kHz Tx at thereceiver site (52◦ N, 8◦ E, great circle distanced = 2210 km) con-verges sufficiently withn = 5 hops.

Ernst D. Schmitter: Remote Auroral Activity Detection 5

Fig. 5. POES polar cap power input for 2010/07/15, model calcula-tion (blue: undisturbed, red: disturbed) and derived reflection heightvariation along propagation path. Polar impact intensity is high ataround 1 UT and again around 14 UT, however only around the firsttime the auroral oval intensity distribution is high enough along thepropagation path (white arrow) to produce significant height reduc-tion - in this case with a maximum of 5.5 km at the transmitter site(bottom line) decreasing slowly along the propagation path.

Fig. 6. For the 3 days 2010/07/14 - 2010/07/16: from top to bottom:Observed signal amplitude, wavelet (Morlet) decomposition withpseudo periods from nearly 0 to 1 hour, sum of absolute values ofthe Morlet coefficients in 3 hour slots, estimated auroral activityindex, ionospheric reflection height reduction (with a maximum of3 km near the transmitter site in this case) along path (bottom lineof height decrease display: Tx, top line: Rx).

Fig. 7. For the 3 days 2010/04/05 - 2010/04/07: POES polar cappower input, Leirvogur magnetogram (horizontal component), ob-served signal, from observed signal estimated activity level, iono-spheric reflection height reduction up to 7 km along propagationpath (bottom: Tx, top: Rx). Special features: SEP onset at first dayshowing up in the day time observations leading to a high activityindex, Tx drop out on the second day which has to be discarded forthe activity level and height variation calculations.

Fig. 8. 14 days period 2010/05/01 - 2010/05/14: POES polarcap power input, Leirvogur magnetogram (horizontal component),north polar auroral activity level estimated from observed VLF/LFsignal .

Fig. 5. POES polar cap power input for 15 July 2010, model cal-culation (blue: undisturbed, red: disturbed) and derived reflectionheight variation along propagation path. Polar impact intensity ishigh at around 01:00 UT and again around 14:00 UT, however onlyaround the first time the auroral oval intensity distribution is highenough along the propagation path (white arrow) to produce signif-icant height reduction – in this case with a maximum of 5.5 km atthe transmitter site (bottom line) decreasing slowly along the prop-agation path.

The width parameterσa depends on the activity levelac andon the time UT letting the small width and low intensity partof the oval point to the sun.

Figure 3 shows as an example the auroral oval intensitydistribution for activity levelac = 9 at 22:00 UT. Indicatedare the 60 and 80 degree latitude circles, the position ofthe geomagnetic pole and the relevant propagation path from(63.9◦ N, 22.5◦ W) to (52◦ N 8◦ E) as white arrow.

www.ann-geophys.net/28/1807/2010/ Ann. Geophys., 28, 1807–1811, 2010

1810 E. D. Schmitter: Remote auroral activity detection






Fig. 6. For the 3 days 14 July 2010–16 July 2010: from top to bot-tom: Observed signal amplitude, wavelet (Morlet) decompositionwith pseudo periods from nearly 0 to 1 h, sum of absolute valuesof the Morlet coefficients in 3 h slots, estimated auroral activity in-dex, ionospheric reflection height reduction (with a maximum of3 km near the transmitter site in this case) along path (bottom lineof height decrease display: Tx, top line: Rx).

3.2 The wave propagation model

Within ray theory the electric field amplitude of a verti-cal Hertzian transmitter dipole is described as the sum of aground wave contribution and multiply reflected sky waves(TM-mode, ordinary ray) followingVolland (1994), Wait(1998), Yoshida et al.(2008):

Ez = AF(d,f,σ )

d+A

n∑j=1

Bj

Lj

eik(Lj −d) (6)

The coefficientA depends on transmitter power and antennaradiation efficiency,F(d,f,σ ) is the ground wave attenua-tion coefficient, depending on great circle distanced, Tx-frequencyf and ground conductivityσ , F = 1 for a per-fectly conducting ground.Lj is the path length of aj -hopwave.k =

2πλ

is the wavenumber. TheBj contain the groundand ionoshperic reflection coefficients. The sky wave pathlengths and the ionospheric reflection coefficients depend onthe effective ionospheric reflection heighth. In the D-layereffective ionospheric reflection is reduced by electron-neutralcollisions which is taken into account via an exponential low-ering of the reflection coefficient with decreasing height. Fig-ure 4 displays the signal amplitude of the 37.5 kHz Tx atthe receiver site (52◦ N, 8◦ E) with d = 2210 km. The resultconverges sufficiently with usingn = 5 hops. The reflectionheight at a specific location along the propagation path varieswith the sun zenith distance anglez. We use:

h = hnight−dhquiet cosm(z)−dhSEP (7)






Fig. 7. For the 3 days 5 April 2010–7 April 2010: POES polarcap power input, Leirvogur magnetogram (horizontal component),observed signal, from observed signal estimated activity level, iono-spheric reflection height reduction up to 7 km along propagationpath (bottom: Tx, top: Rx). Special features: SEP onset at first dayshowing up in the day time observations leading to a high activityindex, Tx drop out on the second day which has to be discarded forthe activity level and height variation calculations.

hnight, dh, m depend on Tx-frequency,Tx-latitude and sea-son. For the NRK/TFK Tx we usedhnight= 86 km,dhquiet=

14 km andm = 0.2 for the summer time for a good reproduc-tion of the observed signal amplitude in a ionospheric quietsituation. Ionization enhancement during solar energetic par-ticle events (SEPs) is modeled by a further height reductiondhSEP.

At given time UT and auroral activity levelac the heightreduction at (lat, long) is modeled with

dhSEP= dh0 lg(I (lat,long,UT,ac)) (8)

for intensitiesI > 1 anddhSEP= 0 otherwise.The coefficientdh0 is calibrated by comparing model

propagation calculations to recorded data. With the auroralintensityI in erg/(s cm2) (to be consistent with NOAA POESdata) we preliminarily founddh0 = 8 km.

Figure5 displays the POES polar cap power input for 15July 2010 (upper trace), our model calculation (2. trace: blue:undisturbed, red: disturbed) and (3. trace) the derived reflec-tion height variation along propagation path. We notice theinfluence of position and extension of the auroral oval: theimpact intensity is high at around 01:00 UT and again around14:00 UT, however only around the first time the auroral ovalpower input is high enough along the propagation path (re-spectively indicated by a white arrow) to produce significantreflection height reduction of the lower ionosphere.

Ann. Geophys., 28, 1807–1811, 2010 www.ann-geophys.net/28/1807/2010/

E. D. Schmitter: Remote auroral activity detection 1811






Fig. 8. 14 days period 1 May 2010–14 May 2010: POES polarcap power input, Leirvogur magnetogram (horizontal component),north polar auroral activity level estimated from observed VLF/LFsignal.

4 Results and conclusions

Referring again to Fig.1 as an example auroral activitycaused by energetic particle precipitation massively influ-ences propagation conditions not only of trans polar prop-agation paths but also paths out of the subpolar domain tolower latitudes during equatorward auroral oval expansions.The daily swing of the precipitation power input oval to-gether with its extension and scale factor quantify the effectin time and intensity. Observations within the first year ofthe new solar cycle show that activity with a total polar capinput reaching and exceeding 30 GW (corresponding to anactivity level of 6.8) safely leaves its fingerprint in the nighttime low frequency signal. During night time two effectscome together along the propagation path: the bulge of theauroral oval is at the transmitter site (the night time maximaalso show up in the Leirvogur magnetograms, see Figs.1, 8)and D-layer attenuation is strongly reduced. However strongsolar energetic particle (SEP) events can be detected in thesignal also during day time as is shown by Fig.7 where aprecipitation onset (shortly exceeding 500 GW input poweraccording to the NOAA data: 5 April 2010 09:07:39 METP-02) takes place day over on the first of the 3 days (5 April2010). The day time precipitation signature can resemblethat of a SID (sudden ionospheric disturbance by X-rays ofa solar flare). Whether the latter was the causing reason ofcourse can be decided quickly by having alook at the NOAAGOES X-ray data for example.

Quantifying the disturbance effects on the signal ampli-tudes wavelet (Morlet) decomposition proved as a reliabletool and taking into account quiet day behavour an estimated(north) auroral activity level can be derived. This informationtogether with an auroral oval model and a propagation pathcalculation is used to estimate the lower ionosphere heightdecrease along the propagation path (cp. Figs.6, 7, 8) thatin turn can be expressed by electron density variations. Theheight decrease along the propagation path also is a proxyfor the mid latitude extension of the auroral electrojet currentsystem. Adapting model calculations to recorded data showsa logarithmic dependence of reflection height decrease withlocal power input intensity according to Eqs. (8). Referringto a frequency of 37.5 kHz we found 8 km decrease for aninput of 10 erg/cm2 s = 10 mJ/m2 s.

A refinement of the models and parameter calibrations issubject of ongoing research.

Acknowledgements.Topical Editor K. Kauristie thanks one anony-mous referee for her/his help in evaluating this paper.

References

Chen, A., Li, J., Yang, G., and Wang, J.: Calculating Auroral OvalPattern by AE Index, Acta Meteorologica Sinica, 22, 91–96,2008.

Cummer, S. A., Bell, T. F., Inan, U. S., and Zanetti, L. J.: VLFremote sensing of the auroral electrojet, J. Geophys. Res. (USA),5381–5389, 1996.

Cummer, S. A., Bell, T. F., Inan, U. S., and Chenette, D. L.: VLFremote sensing of high-energy auroral particle precipitation, J.Geophys. Res. (USA), 7477–7484, 1997.

Kikuchi, T. and Ohtani, A.: IV. Ionospheric Disturbances; 8. Prop-agation Disturbances of VLF Radio Waves on the Trans-PolarPaths, J. Radio Res. Laboratory, 33, special issue No. 1, 203–217, 1986.

Lewalle, J.: Tutorial on Continuous Wavelet Analysis of Experi-mental Data, URL:www.ecs.syr.edu/faculty/lewalle/tutor/tutor.html, Syracuse University, 1995.

Peter, W. B., Chevalier, M., and Inan, U. S.: Subionospheric VLFmeasurements of the effects of geomagnetic storms on the mid-latitude D-region, 11th International Ionosphere Effects Sympo-sium, A070, 3–5 May 2005, pp. 1–8, 2005.

Volland, H.: Longwave Sferics Propagation within the AtmosphericWaveguide, Chapter 3 of Handbook of Atmospheric Electrody-namics, vol. II, edited by: Volland, H., CRC Press, 1995.

Yoshida, M., Yamauchi, T., Horie, T., and Hayakawa, M.: Onthe generation mechanism of terminator times in subiono-spheric VLF/LF propagation and its possible application to seis-mogenic effects, Nat. Hazards Earth Syst. Sci., 8, 129–134,doi:10.5194/nhess-8-129-2008, 2008.

Wait, J. R.: The Ancient and Modern History of EM Ground-WavePropagation, IEEE Antennas and Propagation Magazine, 40(5),7–24, 1998.

www.ann-geophys.net/28/1807/2010/ Ann. Geophys., 28, 1807–1811, 2010

www.ecs.syr.edu/faculty/lewalle/tutor/tutor.html

www.ecs.syr.edu/faculty/lewalle/tutor/tutor.html

remote auroral activity detection and modeling using low ... · 1808 e. d. schmitter: remote...

Documents